Administrators of today's computing environments control the operation of various servers. In order to ensure that these servers provide desired services to users, administrators are often required to be “on call” at night and on weekends. In order to provide administrators with greater flexibility, servers have been adapted for use with services that allow administrators to administer servers remotely. For example, administrators can use a service such as telnet to log into a server remotely.
One problem that arises when administering servers remotely is that, if multiple servers are being administered, each server may have a different remote interface. Each interface may be accessible via a different protocol. Accordingly, in order to be able to administer each of the different servers, an administrator may be forced to install a client for each different protocol on each device that the administrator wants to use for remote administration. For example, an administrator may need to install a proprietary client in order to be able to remotely administer a file server. The same administrator may use a web client in order to be able to remotely administer a database server. If the administrator works with several different servers, the administrator may be forced to frequently switch between interfaces in order to monitor and communicate with each different server. It is desirable to use a single client to remotely administer different servers.
Another problem that arises when administering servers remotely is that it can often take a significant amount of time for an administrator to detect an alert generated by a server. For example, if a server sends alerts via email, the email messages may experience significant delay. Furthermore, an administrator may not notice that she has received an email until some time after the email was sent. Depending on the type of interface used to communicate with a particular server, an administrator may also have to expend a significant amount of effort in order to detect whether any alerts have been generated. For example, even if all of the different servers being administered are administered via a web interface, an administrator still has to install the web interface and then launch the web interface each time the administrator wants to communicate with the servers. This can also increase the time taken to notice an alert (e.g., if an administrator checks for alerts less frequently due to the amount of effort required to log in to the server). Accordingly, it is also desirable to be able to provide alerts in a manner that reduces the amount of administrator effort needed to retrieve those alerts and/or that decreases the length of time likely to elapse before an administrator becomes aware of each alert.